# 404 error: meaning and how to choose a fix

> Tell a retired page from a broken link and decide whether to restore content, redirect it or keep a useful 404 response.


**A 404 error means the server cannot find the requested resource at that address.** A mistyped URL, an old link or a retired page can all produce it. The code alone does not identify the cause or mean that the entire website has stopped working.

As a visitor, check the address and try the site's navigation or search. As the site owner, establish what should exist at that URL before changing its response. A 404 can reveal a problem to repair or be the correct response for content that no longer exists.

## Four cases that need different fixes

Consider a fictional store reviewing its catalog and links:

| Finding | What to establish | Decision if confirmed |
|---|---|---|
| A button links to `/delivery/`, but the page lives at `/shipping/` | The current destination and source link | Correct the link; consider a redirect if the variant receives visits |
| A current page disappeared after a release | The published file, route and configuration | Restore the content at its correct address |
| An article moved to an equivalent URL | The replacement answers the same need | Permanent redirect to the replacement |
| An offer ended with no relevant replacement | Visitors no longer need information retained there | Keep a clear removal response with 404 or 410 |

Prioritize addresses that receive visits, have useful links or sit in purchasing journeys. A log full of random bot-generated paths deserves different attention from a broken booking button. Search Console helps investigate, but an excluded address does not automatically call for new content.

## What is a soft 404?

A page can display “not found” while returning HTTP 200, the normal successful response. Google may interpret this as a *soft 404*: the content looks like an error while the server reports success. [Google's HTTP error documentation](https://developers.google.com/search/docs/crawling-indexing/http-network-errors) explains this treatment.

A well-designed error page does not fix the server status. If the URL should return 404, confirm the code as well as the message. Conversely, a valid page that renders empty needs investigation. Removing it just because a report flagged an error can destroy useful content.

## Help the visitor continue

Explain that the page could not be found and provide real onward routes: a relevant category, site search or a contact option. Keep essential navigation. Avoid telling visitors that everything is fine when the information they wanted is missing.

Sending every unknown address to the homepage can hide broken links and produce irrelevant destinations. Require a defensible replacement for each move. The article on [301 redirects](/en/guides/301-redirect/) shows how to document that relationship.

## Close the actual issue

Record the exact URL, how visitors reach it, the observed cause and the expected outcome. After a repair, follow the link from its source and check the first response and final content. Update internal links and sitemap entries when public URLs change. Close the review when the link reaches the expected content or the retired URL returns the intended status.

## Plan your next post in HeyMark.

Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.

[Start free](https://app.heymark.ai) · [See how it works](https://heymark.ai/en/#product)

## Developer and agent resources

- [HeyMark MCP documentation](https://heymark.ai/en/mcp/)
- [llms.txt](https://heymark.ai/llms.txt)
- [Full site content for language models](https://heymark.ai/llms-full.txt)
- MCP protocol endpoint: `POST https://mcp.heymark.ai`
- OAuth protected-resource metadata: [/.well-known/oauth-protected-resource](https://mcp.heymark.ai/.well-known/oauth-protected-resource)
- MCP server card: [/server-card](https://mcp.heymark.ai/server-card)
